Light-Weight Distributed Real-Time System Node Status Probe And Study Mechanism

Distributed real-time system technique currently has been applied in many working systems, such as real-time stock system, air-line management system and military systems. Only by guaranteeing the coordination of system nodes can these systems achieve their system goals successfully. To keep a distributed real-time system’s nodes in the same pace, communications among these nodes should follow the rules of low delay and high reliability. Therefore, a fast and reliable mechanism of nodes probing and synchronizing is necessary to distributed real-time systems.Some distributed real-time systems are running on special private network. Every node in special private network demands highly autonomy. The infrastructure of special private network also has different kinds of implement. However, some current node status probe and study mechanisms seriously rely on the network center and some have high communication cost by using long message. On the other side, these mechanisms only emphasize the protocol communication between network nodes and do not take the inside-node units’ status into the consideration of node status probe and study. Because of all the reasons above, for the nodes probing and synchronizing problems of distributed real-time systems, we should build a node status probe and study mechanism, which is suitable for none-center network, light-weight and low-cost, to meet the needs of probe between network nodes and discovery of units inside a network node. A middleware based on this mechanism is also needed. This paper introduces the relevant concepts and techniques about distributed real-time system. After trying to analyze the technical features of current node status probe and study mechanisms, we put forward the major study goals of this paper. Next, this paper focuses on the design of a node status probe and study mechanism, emphasize on both the probe between nodes and the discovery of inside-node units. Specifically, we raise a protocol design based on UDP and a node unit status sharing mechanism using shared memory technique. Based on this mechanism, the following part of this paper shows how to design and implement the according middleware to meet the requirements of distributed real-time system nodes’probing and synchronizing.